Back on 17th June I saw this alarming message for the first time from my new C-8:
Corrupt database found
Hub cannot continue.
Went through the advised steps and I could recover a recent local backup, which got the thing up and running again. All has been fine until I was away and could no longer control anything. On my return, the same message.
Once again I've restored from a local backup and it's working - but this is twice now and I'm beginning to be suspicious. I've never experienced this on any Hubitat hub before and the C-7 which sits next to it (and shares the same more-than-powerful-enough three-output adapter) hasn't missed a beat, ever.
While you wait for a response from support, do you have any details on what you are running, e.g. major device types or apps/integrations you are running?
May also be worth mentioning which platform version you are running, e.g. 2.3.X.X.
Based on your reference to the power source and previous advice, I'm assuming you have read through this thread:
The C-7 and the C-8 run purely Zigbee devices and all of the same apps. Third party are CoCoHue, Homerbridge, Hubitat Package Manager, Influx DB Logger. The rest are all Hubitat, nothing "exciting". The C-7 does all of my AlertMe / Iris V1 gear, the C-8 does everything else - though that's really only some Samotech and Tuya switch modules and Aurora dimmers, only 24 devices on the C-8. All of this ran perfectly for over a year on the C-7 I migrated from to get to the C-8.
They also both run the latest 2.3.5.152 firmware.
My main concern is whether this C-8 has some hardware fault, possibly storage-related, that is leading to the corruptions. And yes, I did indeed read that thread, which is why I checked that the power was stable.
Unfortunately I have seen issues with all of those custom apps bringing hubs down if they are ill formatted. Your best course of action is screening your Logs (past logs) for any errors. Also the Apps and Devices tabs on Logs page may give you some clues. If nothing obvious, I'd say, to temporarily disable the custom apps and see if that helps. Soft Reset when the database becomes corrupted is your temporary solution to bring your hub back online, but is not the cure.
Could the migration have created an error in those apps? The configuration was brought over and hasn’t been changed since, so far as I remember.
Sadly I can’t remove them for potentially months at a time as they’re essential to my use of the hub, though I’ll certainly check for errors in the logs and set them back up if needs be.
I’ll report back on the past logs check, I’m not in a position to check that this minute.
I've employed my usual scorched earth policy, removed all of the third party apps, reinstalled and set them up again. Reconfigured the button rules and everything is working correctly.
Interestingly, there were no errors in the log before doing this.
Is there somewhere I can see whether the hub has been rebooting? My two hubs are run from the same power supply which provides multiple outputs; the only thing to change has been the swap to a C-8 from a C-7 and the C-7 still in use continues to run without fault, so the power supply seems like an unlikely candidate for fault in this situation.
I'm not saying it isn't that, and if problems continue it'll be my next troubleshooting step.
Sorry, is this a Rule Machine thing? I've got a notification method I can use, I just have no idea where I'm looking for "system startup" to be a trigger event.
Thanks for the pointers, @rlithgow1 - I'll get that set up and keep an eye on things. Logging into InfluDB shows no interruptions so far and everything is behaving at the moment.
So I've been experiencing random, about weekly, database corruption events requiring a soft reset on only one of my hubs... @bobbyD or @gopher.ny can you take a look at the engineering logs and see if you can tell why?
The latest one was sometime 0043 hours this morning. The hub's name is Main...
